
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import warnings
warnings.filterwarnings("ignore")

import plotly.express as px
from plotly.offline import download_plotlyjs, init_notebook_mode, plot, iplot
from plotly import graph_objs as go


df = pd.read_csv('./data/world_population_by_country.csv')
df


df.shape

(234, 19)


df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 234 entries, 0 to 233
Data columns (total 19 columns):
 #   Column           Non-Null Count  Dtype  
---  ------           --------------  -----  
 0   country          234 non-null    object 
 1   rank             234 non-null    int64  
 2   area             234 non-null    float64
 3   landAreaKm       234 non-null    float64
 4   cca2             233 non-null    object 
 5   cca3             234 non-null    object 
 6   netChange        226 non-null    float64
 7   growthRate       234 non-null    float64
 8   worldPercentage  228 non-null    float64
 9   density          234 non-null    float64
 10  densityMi        234 non-null    float64
 11  place            234 non-null    int64  
 12  pop1980          234 non-null    int64  
 13  pop2000          234 non-null    int64  
 14  pop2010          234 non-null    int64  
 15  pop2022          234 non-null    int64  
 16  pop2023          234 non-null    int64  
 17  pop2030          234 non-null    int64  
 18  pop2050          234 non-null    int64  
dtypes: float64(7), int64(9), object(3)
memory usage: 34.9+ KB


df.describe().T.sort_values("50%", ascending = False).style.background_gradient(cmap = "RdPu")\
    .bar(subset = ["mean"], color = "red").bar(subset = ["max"], color = "green")


df.isna().sum()

country            0
rank               0
area               0
landAreaKm         0
cca2               1
cca3               0
netChange          8
growthRate         0
worldPercentage    6
density            0
densityMi          0
place              0
pop1980            0
pop2000            0
pop2010            0
pop2022            0
pop2023            0
pop2030            0
pop2050            0
dtype: int64


# Distribution du taux de croissance de la population
plt.figure(figsize=(10, 6))
sns.histplot(data=df, x='growthRate', bins=20, kde=True)
plt.title('Population Growth Rate Distribution')
plt.xlabel('Growth Rate (%)')
plt.ylabel('Count')
plt.show()


annee_23 = df.sort_values(by='pop2023', ascending=False)

fig1 = px.treemap(annee_23, path=[annee_23.country[:10]], values = annee_23['pop2023'][:10], title="Top 10 des pays les plus peuplés en 2023")
fig1.show()


fig2 = px.bar(annee_23,x=annee_23.country[:10], y= annee_23['pop2023'][:10], color=annee_23.country[:10])
fig2.update_layout(xaxis_title = "Countries", yaxis_title="Population of countries")
fig2.show()

fig = px.choropleth(locations = annee_23.cca3, color=annee_23['pop2023'],
                    color_continuous_scale=px.colors.sequential.Rainbow,
                    title='population mondiale en 2023')
fig.show()


#  Les 10 pays les plus peuplés en 2023
top_pop_2023 = df.nlargest(10, "pop2023")

# Top 10 des pays ayant la plus grande superficie
top_area = df.nlargest(10, "area")

# Top 10 des pays ayant la plus forte densité de population en 2023
top_density_2023 = df.nlargest(10, "density")

fig, ax = plt.subplots(3, 1, figsize=(15, 20))

ax[0].barh(top_pop_2023["country"], top_pop_2023["pop2023"], color="green")
ax[0].invert_yaxis()
ax[0].set_xlabel("Poupulation en 2023")
ax[0].set_title("Les 10 pays les plus peuplés en 2023")

ax[1].barh(top_area["country"], top_area["area"], color='blue')
ax[1].invert_yaxis()
ax[1].set_xlabel("Area(km2)")
ax[1].set_title("Top 10 des pays ayant la plus grande superficie")

ax[2].barh(top_density_2023["country"], top_density_2023["density"], color='skyblue')
ax[2].invert_yaxis()
ax[2].set_xlabel("Densité de la population en 2023")
ax[2].set_title("Top 10 des pays ayant la plus forte densité de population en 2023")

plt.tight_layout()
plt.show()


# Histogrammes de certaines colonnes numériques
df[['area', 'pop2023', 'density']].hist(bins=10, figsize=(15, 10), color='skyblue', edgecolor='black')
plt.tight_layout()
plt.show()


# Evolution de la population de 1980 à 2023 pour chaque pays
df["pop_change_1980_2023"] = df["pop2023"] - df["pop1980"]

# Top 10 des pays avec la plus forte augmentation de population
top_pop_increase = df.nlargest(10, "pop_change_1980_2023")

# Top 10 des pays ayant la plus forte diminution de population
top_pop_decrease = df.nsmallest(10, "pop_change_1980_2023")

fig, ax = plt.subplots(2,1, figsize=(15,20))

ax[0].barh(top_pop_increase["country"], top_pop_increase["pop_change_1980_2023"], color="red")
ax[0].invert_yaxis()
ax[0].set_xlabel("Évolution de la population de 1980 à 2023")
ax[0].set_title("Top 10 des pays ayant la plus forte augmentation de population de 1980 à 2023 ")


ax[1].barh(top_pop_decrease["country"], top_pop_decrease["pop_change_1980_2023"], color="brown")
ax[1].invert_yaxis()
ax[1].set_xlabel("Évolution de la population de 1980 à 2023")
ax[1].set_title("Top 10 des pays ayant connu la plus forte baisse de population entre 1980 et 2023 ")

plt.tight_layout()
plt.show()


# Information sur le Canada
canada_data = df[df['country'] == 'Canada']
canada_data.transpose()


# Information sur la France
France_data = df[df['country'] == 'France']
France_data.transpose()


# Evolution de la population par superficie pour chaque pays
df['pop_change_per_area'] = df['pop_change_1980_2023'] / df['landAreaKm']

#  Top 10 des pays où la population a le plus changé par rapport à la superficie du territoire
top_pop_change_per_area = df.nlargest(10, 'pop_change_per_area')

fig, ax = plt.subplots(figsize=(10, 8))

ax.barh(top_pop_change_per_area['country'], top_pop_change_per_area['pop_change_per_area'], color='skyblue')
ax.invert_yaxis()
ax.set_xlabel('Évolution de la population par territoire de 1980 à 2023 (personnes par km2)')
ax.set_title('Top 10 pays évolution de la population par superficie entre 1980 et 2023')

plt.tight_layout()
plt.show()


# densité de population pour chaque année et pour chaque pays
df['density_1980'] = df['pop1980'] / df['landAreaKm']
df['density_2000'] = df['pop2000'] / df['landAreaKm']
df['density_2010'] = df['pop2010'] / df['landAreaKm']
df['density_2022'] = df['pop2022'] / df['landAreaKm']

# Top 10 pays ayant la plus forte densité de population en 1980, 2000, 2010, 2022
top_density_1980 = df.nlargest(10, 'density_1980')
top_density_2000 = df.nlargest(10, 'density_2000')
top_density_2010 = df.nlargest(10, 'density_2010')
top_density_2022 = df.nlargest(10, 'density_2022')

# Plotting
fig, ax = plt.subplots(2, 2, figsize=(20, 15))

ax[0, 0].barh(top_density_1980['country'], top_density_1980['density_1980'], color='skyblue')
ax[0, 0].invert_yaxis()
ax[0, 0].set_xlabel('Densité de la population en 1980 (people per km²)')
ax[0, 0].set_title('Top 10 pays par densité de population en 1980')

ax[0, 1].barh(top_density_2000['country'], top_density_2000['density_2000'], color='skyblue')
ax[0, 1].invert_yaxis()
ax[0, 1].set_xlabel('Densité de la population en 2000 (people per km²)')
ax[0, 1].set_title('Top 10 pays par densité de population en 2000')

ax[1, 0].barh(top_density_2010['country'], top_density_2010['density_2010'], color='skyblue')
ax[1, 0].invert_yaxis()
ax[1, 0].set_xlabel('Densité de la population en 2010 (people per km²)')
ax[1, 0].set_title('Top 10 pays par densité de population en2010')

ax[1, 1].barh(top_density_2022['country'], top_density_2022['density_2022'], color='skyblue')
ax[1, 1].invert_yaxis()
ax[1, 1].set_xlabel('Densité de la population en 2022 (people per km²)')
ax[1, 1].set_title('Top 10 pays par densité de population en 2022')

plt.tight_layout()
plt.show()


# matrice de correlation 
correlation_matrix = df[['area', 'landAreaKm', 'netChange', 'growthRate', 'density', 'pop1980', 'pop2000', 'pop2010', 'pop2022', 'pop2023']].corr()

plt.figure(figsize=(12, 8))

# Générer un masque pour le triangle supérieur
mask = np.triu(np.ones_like(correlation_matrix, dtype=bool))

#  heatmap with avec le masque et le bon rapport hauteur/largeur
sns.heatmap(correlation_matrix, mask=mask, cmap='viridis', vmax=1, vmin=-1,
            square=True, linewidths=.5, cbar_kws={"shrink": .5}, annot=True)

plt.title('Correlation Matrix Heatmap')
plt.show()


# Tendances dans le temps pour un sous-ensemble de pays
subset_countries = ['China', 'India', 'United States', 'Indonesia', 'Pakistan', 'Brazil', 'Nigeria', 'Bangladesh', 'Russia', 'Mexico']

subset_df = df[df['country'].isin(subset_countries)]
subset_df = subset_df[['country', 'pop1980', 'pop2000', 'pop2010', 'pop2022', 'pop2023']]

# Transformation en format long
subset_df_melted = subset_df.melt(id_vars='country', var_name='year', value_name='population')

# Conversion
subset_df_melted['year'] = subset_df_melted['year'].str[3:].astype(int)

# Plotting
plt.figure(figsize=(15, 10))
sns.lineplot(data=subset_df_melted, x='year', y='population', hue='country')
plt.title('Tendances démographiques dans le temps pour certains pays')
plt.xlabel('Year')
plt.ylabel('Population')

plt.show()


# diagramme à barres montrant le taux de croissance de la population des 10 pays les plus peuplés en 2023.

top10_populous_2023 = df.nlargest(10, 'pop2023')['country']

#  taux de croissance de la population pour ces pays
growth_rate_top10 = df[df['country'].isin(top10_populous_2023)][['country', 'growthRate']]

# Trier les pays par taux de croissance
growth_rate_top10 = growth_rate_top10.sort_values('growthRate', ascending=False)

# Plotting
plt.figure(figsize=(12, 6))
sns.barplot(x='growthRate', y='country', data=growth_rate_top10, palette='viridis')
plt.title('Population Growth Rate for Top 10 Most Populous Countries in 2023')
plt.xlabel('Population Growth Rate')
plt.ylabel('Country')
plt.grid(True)
plt.show()


df_clean = df.dropna(subset=['netChange', 'worldPercentage', 'cca2'])


from sklearn.cluster import KMeans
from sklearn.preprocessing import StandardScaler

# Select features for clustering
features = ['area', 'landAreaKm', 'netChange', 'growthRate', 'density', 'pop1980', 'pop2000', 'pop2010', 'pop2022', 'pop2023']

# Standardize the features
scaler = StandardScaler()
scaled_features = scaler.fit_transform(df_clean[features])

# Perform KMeans clustering
kmeans = KMeans(n_clusters=5, random_state=42)
kmeans.fit(scaled_features)

# Add cluster labels to the dataframe
df_clean['cluster'] = kmeans.labels_

# Display the size of each cluster
df_clean['cluster'].value_counts().sort_index()

0    142
1      2
2      5
3      2
4     73
Name: cluster, dtype: int64


# boxplots pour visualiser les valeurs aberrantes.
fig, ax = plt.subplots(2, 5, figsize=(20, 10))

for i, feature in enumerate(features):
    sns.boxplot(x=df[feature], ax=ax[i//5, i%5])
    ax[i//5, i%5].set_title(feature)

plt.tight_layout()
plt.show()


# valeurs moyennes pour chaque groupe
cluster_means = df_clean.groupby('cluster')[features].mean()

# Affichage des valeurs moyennes des clusters 
cluster_means.transpose()


# population totale de chaque groupe en 2023
cluster_pop2023 = df_clean.groupby('cluster')['pop2023'].sum()

# Plotting
plt.figure(figsize=(10, 6))
cluster_pop2023.sort_index().plot(kind='bar', stacked=True, color=sns.color_palette('viridis', 5))
plt.title('Population totale de chaque groupe en 2023')
plt.xlabel('Cluster')
plt.ylabel('Population Totale')
plt.show()


# 5. Bubble Chart of Population vs Area, colored by Growth Rate
plt.figure(figsize=(12, 8))
sns.scatterplot(x='area', y='pop2023', size='growthRate', sizes=(20, 200), hue='growthRate', palette='viridis', data=df_clean)
plt.title('Relationship Between Area, Population, and Population Growth Rate in 2023')
plt.xlabel('Area (km²)')
plt.ylabel('Population in 2023')
plt.grid(True)
plt.legend(bbox_to_anchor=(1.01, 1),borderaxespad=0)
plt.show()

	country	rank	area	landAreaKm	cca2	cca3	netChange	growthRate	worldPercentage	density	densityMi	place	pop1980	pop2000	pop2010	pop2022	pop2023	pop2030	pop2050
0	India	1	3287590.00	2973190.00	IN	IND	0.4184	0.0081	0.1785	480.5033	1244.5036	356	696828385	1059633675	1240613620	1417173173	1428627663	1514994080	1670490596
1	China	2	9706961.00	9424702.90	CN	CHN	-0.0113	-0.0002	0.1781	151.2696	391.7884	156	982372466	1264099069	1348191368	1425887337	1425671352	1415605906	1312636325
2	United States	3	9372610.00	9147420.00	US	USA	0.0581	0.0050	0.0425	37.1686	96.2666	840	223140018	282398554	311182845	338289857	339996563	352162301	375391963
3	Indonesia	4	1904569.00	1877519.00	ID	IDN	0.0727	0.0074	0.0347	147.8196	382.8528	360	148177096	214072421	244016173	275501339	277534122	292150100	317225213
4	Pakistan	5	881912.00	770880.00	PK	PAK	0.1495	0.0198	0.0300	311.9625	807.9829	586	80624057	154369924	194454498	235824862	240485658	274029836	367808468
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
229	Montserrat	230	102.00	102.00	MS	MSR	NaN	-0.0009	NaN	43.0000	111.3700	500	11452	5138	4938	4390	4386	4301	3781
230	Falkland Islands	231	12173.00	12173.00	FK	FLK	NaN	0.0029	NaN	0.3114	0.8066	238	2240	3080	3187	3780	3791	3869	3779
231	Niue	232	261.00	261.00	NU	NIU	0.0000	0.0005	NaN	7.4138	19.2017	570	3637	2074	1812	1934	1935	1948	2096
232	Tokelau	233	12.00	10.00	TK	TKL	NaN	0.0118	NaN	189.3000	490.2870	772	1647	1666	1367	1871	1893	2046	2430
233	Vatican City	234	0.44	0.44	VA	VAT	NaN	0.0157	NaN	1177.2727	3049.1364	336	733	651	596	510	518	561	731

	count	mean	std	min	25%	50%	75%	max
pop2050	234.000000	41486278.790598	148167567.122402	731.000000	546605.750000	6352397.000000	35686137.750000	1670490596.000000
pop2030	234.000000	36514605.333333	141782710.184894	561.000000	456149.000000	6178231.000000	26163111.500000	1514994080.000000
pop2023	234.000000	34374424.743590	137386405.597263	518.000000	422598.250000	5643895.000000	23245367.250000	1428627663.000000
pop2022	234.000000	34074414.713675	136766424.804728	510.000000	419738.500000	5559944.500000	22476504.750000	1425887337.000000
pop2010	234.000000	29845235.042735	124218487.631581	596.000000	393149.000000	4942770.500000	19159567.500000	1348191368.000000
pop2000	234.000000	26269468.816239	111698206.719070	651.000000	327242.000000	4292907.000000	15762301.000000	1264099069.000000
pop1980	234.000000	18984616.982906	81785186.081872	733.000000	229614.250000	3141145.500000	9826053.750000	982372466.000000
area	234.000000	581449.983590	1761840.665609	0.440000	2650.000000	81199.500000	430425.750000	17098242.000000
landAreaKm	234.000000	557112.276239	1689971.526445	0.440000	2625.875000	75689.250000	404787.625000	16376870.000000
place	234.000000	439.085470	253.295484	4.000000	223.000000	439.000000	659.750000	894.000000
densityMi	234.000000	1168.836388	5126.548664	0.357400	102.946450	252.475800	629.185350	55433.006400
rank	234.000000	117.500000	67.694165	1.000000	59.250000	117.500000	175.750000	234.000000
density	234.000000	451.288182	1979.362419	0.138000	39.747650	97.481000	242.928650	21402.705200
growthRate	234.000000	0.009737	0.012350	-0.074500	0.002325	0.008200	0.016850	0.049800
netChange	226.000000	0.010306	0.034774	-0.028600	0.000000	0.000900	0.008000	0.418400
worldPercentage	228.000000	0.004407	0.017375	0.000000	0.000100	0.000750	0.002925	0.178500

cluster	0	1	2	3	4
area	1.849054e+05	6.497276e+06	1.053266e+07	17.46000	5.932312e+05
landAreaKm	1.690771e+05	6.198946e+06	1.010801e+07	17.45000	5.832978e+05
netChange	1.660563e-03	2.035500e-01	2.006000e-02	0.00015	2.170548e-02
growthRate	3.612676e-03	3.950000e-03	5.360000e-03	0.00410	2.300822e-02
density	3.705139e+02	3.158865e+02	1.592900e+01	19775.60260	1.402950e+02
pop1980	8.545361e+06	8.396004e+08	1.045807e+08	136204.00000	1.404854e+07
pop2000	1.063570e+07	1.161866e+09	1.309637e+08	232180.50000	2.268328e+07
pop2010	1.149305e+07	1.294402e+09	1.413523e+08	295237.50000	2.812929e+07
pop2022	1.239412e+07	1.421530e+09	1.525897e+08	365818.50000	3.567204e+07
pop2023	1.241911e+07	1.427150e+09	1.532168e+08	370223.00000	3.638757e+07

La donnée est l’actif stratégique de la révolution numérique

World Population Analysis

Commençons par examiner les pays ayant la plus grande population, la plus grande superficie, et la densité de population la plus élevée.¶

Qu'en est il du canada et de la France ?¶

Correlations¶

Analyse des tendances temporelles¶

Taux de croissance de la population pour les 10 pays les plus peuplés en 2023¶

Nous allons maintenant passer à l'analyse de regroupement des pays. Pour cela, nous allons utiliser une technique de clustering, telle que le K-means clustering, pour regrouper les pays en fonction de caractéristiques similaires.¶

	37
country	Canada
rank	38
area	9984670.0
landAreaKm	8965590.0
cca2	CA
cca3	CAN
netChange	0.0104
growthRate	0.0085
worldPercentage	0.0048
density	4.3256
densityMi	11.2032
place	124
pop1980	24511510
pop2000	30683313
pop2010	33963412
pop2022	38454327
pop2023	38781291
pop2030	41008596
pop2050	45890819
pop_change_1980_2023	14269781

	22
country	France
rank	23
area	551695.0
landAreaKm	547557.0
cca2	FR
cca3	FRA
netChange	0.004
growthRate	0.002
worldPercentage	0.0081
density	118.2646
densityMi	306.3052
place	250
pop1980	53713830
pop2000	58665453
pop2010	62444567
pop2022	64626628
pop2023	64756584
pop2030	65543452
pop2050	65827072
pop_change_1980_2023	11042754